Journal
PHYSICS IN MEDICINE AND BIOLOGY
Volume 59, Issue 21, Pages 6637-6657Publisher
IOP PUBLISHING LTD
DOI: 10.1088/0031-9155/59/21/6637
Keywords
computed tomography; virtual lesions; image quality; simulation
Funding
- GE healthcare
Ask authors/readers for more resources
Realistic three-dimensional (3D) mathematical models of subtle lesions are essential for many computed tomography (CT) studies focused on performance evaluation and optimization. In this paper, we develop a generic mathematical framework that describes the 3D size, shape, contrast, and contrast-profile characteristics of a lesion, as well as a method to create lesion models based on CT data of real lesions. Further, we implemented a technique to insert the lesion models into CT images in order to create hybrid CT datasets. This framework was used to create a library of realistic lesion models and corresponding hybrid CT images. The goodness of fit of the models was assessed using the coefficient of determination (R-2) and the visual appearance of the hybrid images was assessed with an observer study using images of both real and simulated lesions and receiver operator characteristic (ROC) analysis. The average R-2 of the lesion models was 0.80, implying that the models provide a good fit to real lesion data. The area under the ROC curve was 0.55, implying that the observers could not readily distinguish between real and simulated lesions. Therefore, we conclude that the lesion-modeling framework presented in this paper can be used to create realistic lesion models and hybrid CT images. These models could be instrumental in performance evaluation and optimization of novel CT systems.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available